When I capture from Source, I use Source Recorder. I tend to have problems with skipped frames when I use FRAPS, and there's very little control over how audio is captured. With Source Recorder, I first record a demo because the 'startmovie' process tends to get really bogged down. Source Recorder (by default) outputs each individual frame as TGA, and the audio as raw WAV. I put the individual frames into VirtualDub and add the WAV audio, then set the framerate, and finally export as AVI with no compression. I edit the raw video in Vegas (sometimes), export to (again) raw AVI,
then encode using 10000kBps 2-pass x264 along with 192kBps AAC (VBR) in
MeGUI. I use MKV as the container format.
here's an example of a 1280x720 (not displayed as, though) video on youtube.
the video I uploaded was literally
10.1MB. yes, it's only 32 seconds, but at 1280x720...
furthermore, the raw video was around
2.2GB. %0.0045 compression ratio.